BUG: groupby().rolling("freq") validation raises even if dates are sorted within each group

[JILPulvino]

Pandas version checks

• I have checked that this issue has not already been reported.

• I have confirmed this bug exists on the latest version of pandas.

• I have confirmed this bug exists on the main branch of pandas.

Reproducible Example

import pandas as pd
print(pd.__version__)
test_df = pd.DataFrame({
    'Name':['L', 'L', 'L',
                    'M', 'M', 'M', 'M', 'M', 'M'],
    'ID': ['L1', 'L2', 'L3',
              'M1', 'M2', 'M3', 'M4', 'M5', 'M6'],
    'Date': [
        '2005-01-01', '2006-01-01', '2013-01-01', 
        '2005-01-01', '2006-01-01', '2006-01-01', '2017-01-01', '2006-02-01', '2006-03-01'],
    'indicator': [0, 1, 0,
                          0, 1, 0, 0, 1, 0],
    'rolling_sum':[0, 0, 1,
                              0, 0, 0, 0, 1, 2]})
test_df['Date'] = pd.to_datetime(test_df['Date'], errors='coerce')
test_df.groupby('Name').rolling('731D', on='Date')['indicator'].sum().reset_index()

Issue Description

Grouping by a column and then rolling over the grouped dataframe results in [ValueError: Date must be monotonic](). Ultimately, the operation should be able to produce the rolling_sum column in the example.

Expected Behavior

This is not an issue in pandas 1.3.5 but is an issue in 1.4 and higher. I expect this function to create a grouped dataframe and to then perform the rolling summation on the provided column using the date window provided.

Installed Versions

INSTALLED VERSIONS

commit : 06d2301
python : 3.8.12.final.0
python-bits : 64
OS : Darwin
OS-release : 21.1.0
Version : Darwin Kernel Version 21.1.0: Wed Oct 13 17:33:24 PDT 2021; root:xnu-8019.41.5~1/RELEASE_ARM64_T8101
machine : x86_64
processor : i386
byteorder : little
LC_ALL : None
LANG : None
LOCALE : None.UTF-8

pandas : 1.4.1
numpy : 1.22.2
pytz : 2021.3
dateutil : 2.8.2
pip : 22.0.3
setuptools : 60.6.0
Cython : None
pytest : None
hypothesis : None
sphinx : None
blosc : None
feather : None
xlsxwriter : None
lxml.etree : None
html5lib : None
pymysql : None
psycopg2 : None
jinja2 : 3.0.3
IPython : 8.0.1
pandas_datareader: None
bs4 : None
bottleneck : None
fastparquet : None
fsspec : None
gcsfs : None
matplotlib : None
numba : None
numexpr : None
odfpy : None
openpyxl : 3.0.9
pandas_gbq : None
pyarrow : 6.0.1
pyreadstat : None
pyxlsb : None
s3fs : None
scipy : None
sqlalchemy : None
tables : None
tabulate : None
xarray : None
xlrd : None
xlwt : None
zstandard : None

[phofl]

Hi, thanks for your report. Could you please trim your example down a bit? It should not contain more values than strictly necessary.

[mroeschke]

Thanks for the report.

This behavior was a bug fix in 1.4 as seen in https://pandas.pydata.org/docs/whatsnew/v1.4.0.html#groupby-resample-rolling

Bug in Groupby.rolling() when non-monotonic data passed, fails to correctly raise ValueError (GH43909)

But even if the dates are sorted with the groups this error still raised which seems like a bug

In [2]: test_df.sort_values(["Name", "Date"]).groupby('Name').rolling('2000D', on='Date')
ValueError: Date must be monotonic

[simonjayhawkins]

#46065 reverted, #46078

[simonjayhawkins]

But even if the dates are sorted with the groups this error still raised which seems like a bug

since this worked (for the sorted groups case) in 1.3.5, labelling as a regression.

first bad commit: [b1827ea] BUG: 43909 - check monoticity of rolling groupby (#44068)